Convergence for inviscid flow over sylinder
Hello,
I have a sylinder which I send an incompressible, inviscid flow over. All walls and sylinder is set to SLIPcondition. First I want to send water over the sylinder with, for instance, 10 m/s, and afterwards increase this to 10.5, and compare the total kinetic energy of the two controlvolumes. But flow over a sylinder isn't steady, is it? I've got great problems with convergence using steady flow. The most important thing is that I compare the two cases for flow at the exact same moment. Can I run a transient analysis for this? Or is there a good way to get convergence using steady flow? The reason for me wanting to do this, is that I want to find the "addedmass" of my object. Optimally I would want an irrotational flow (potential), but I guess this is not possible with my CFD package (CdAdapco StarCD.) Thanks in advance. 
Re: Convergence for inviscid flow over sylinder
Incompressible, inviscid, rotational (Euler) flow over a cylinder is problematic. There may be a steady solution, but it would be very difficult to obtain numerically for this reason: You cannot avoid numerical dissipation. Numerical dissipation will act somewhat (in a very uncontrolled and arbitrary way) as if you had a small nonzero viscosity. This will result in an effective Reynolds number which may be very large (also depending on free stream velocity and cylinder size). Large Reynolds number means: No steady solution! Small numerical dissipation will really spoil it for you. There could be two ways out of this dilemna: a) make sure you choose a small velocity over a large cylinder diameter, b) increase your artificial dissipation to get a lower effective Reynolds number (this will have a damping effect on the instability). Either way it's going to be an "engineered" solution that depends a lot on numerical dissipation.
Here is what you really should do: Run a potential flow code or apply the wellknown potential solution analytically. 
Re: Convergence for inviscid flow over sylinder
When selecting your viscous model in your code, select the inviscous model (commercial codes have this option). If the code you are using does not have that option, you can always select a viscous model and entering a value of Zero for the viscosity coefficient

All times are GMT 4. The time now is 15:33. 